Arkadiusz Durasiewicz

One of the main challenges in higher education management is the complexity of resource optimization and increasing volumes of data, which limits the efficiency and accuracy of decision-making. The application of artificial intelligence can address these issues.
The present study aims to identify the key trends, knowledge gaps, and opportunities for further research into the economic effects of using artificial intelligence and ChatGPT tools in higher education. For this purpose, a systematic literature review was conducted to identify and screen the scientific articles related to the topic of this study indexed in Web of Science and Scopus from 1986 to 2024. A total of 234 articles were selected, all demonstrating positive growth both in scholarly output and citation count. The study identified the key contributors to scientific research on this topic by region (the United States, China, and India). It concluded that the relevant research centers are still at an early stage of their development. Based on bibliometric clusters formed by co-occurrence relations, three main areas of research were defined: 1) artificial intelligence in education for decision-making; 2) process automation and digital transformation in educational institutions; 3) artificial intelligence technologies and their application in education.
The study highlights the main areas of economic effects of artificial intelligence and ChatGPT tools in higher education, including reducing administrative costs, saving time for teachers and students, and improving the quality and accessibility of educational process.

Acknowledgments
The publication is part of the research topic “Economic Basics of Technology Diffusion into the National Economy of Ukraine Considering Best International Practices” (№0124U003482).

Type of the article: Research Article

Abstract
This study examines whether logistics performance mediates the relationship between economic complexity and national innovation outcomes. The sample comprises the top-20 economies of the Global Innovation Index (2025) and Ukraine, observed over 2020–2024. The mediation model is estimated on the 2024 cross-section (n = 21) using OLS and the Baron–Kenny procedure. Cross-sectional point estimates indicate an indirect effect via the Logistics Performance Index of 7.585 GII points, 46.3% of the total effect (c = 16.366, p < 0.001), while the software–expenditure channel fails the Baron–Kenny conditions. The pooled panel corroborates the channel: the indirect effect equals 6.551 points (52.0% of the total effect), with the LPI–GII path highly significant (p < 0.001). Sensitivity analysis shows that the mediation is identified primarily by the contrast between the innovation frontier and Ukraine: excluding Ukraine, the ECI–LPI path loses significance while the LPI–GII path remains robust, consistent with saturation of the logistics channel within the frontier, where LPI varies only between 3.6 and 4.3. The bootstrap confidence interval for the indirect effect includes zero at n = 21, so the mediation findings are suggestive rather than confirmatory. Fixed-effects estimation shows that the ECI–GII relationship is predominantly structural (within-R2 = 0.056), while LPI retains within-country significance (β = 3.60, p = 0.042). An illustrative arithmetic scenario translates Ukraine’s LPI gap to the sample median into approximately 16 GII points. Findings position logistics infrastructure as a first-order margin for catching-up economies seeking to convert productive complexity into innovation capacity.

Acknowledgments
This research contains results of the research “Fundamentals of Sustainable and Inclusive Regional Spatial Development for Post-War Reconstruction in the Context of Digital Transformation” (№ 0125U001620, 2025-2027) funded by a grant from the state budget of Ukraine.